Skip to main content

Table 4 Validity estimates (95% confidence intervals) for the cardiovascular mortality algorithm, overall and by site

From: Validity of an algorithm to identify cardiovascular deaths from administrative health records: a multi-database population-based cohort study

Measure

All Sites

Canadian Sites

CPRD

All Canadian Sites

AB

BC

MB

ON

QC

Sensitivity

64.8 (63.6, 66.0)

63.3 (62.1, 64.5)

83.1 (78.9, 86.6)

64.3 (60.6, 67.7)

54.8 (49.9, 59.6)

63.4 (61.6, 65.2)

60.4 (58.1, 62.7)

87.3 (83.6, 90.3)

Specificity

74.9 (74.1, 75.6)

76.3 (75.6, 77.0)

67.5 (64.0, 70.8)

66.9 (65.0, 68.6)

75.6 (70.8, 80.0)

77.2 (76.0, 78.4)

82.4 (81.5, 83.7)

53.2 (49.8, 56.5)

PPV

54.5 (53.7, 55.3)

55.4 (54.5, 56.3)

57.0 (52.7, 61.1)

34.0 (31.4, 36.5)

72.8 (67.5, 77.5)

61.6 (60.2, 63.0)

57.6 (55.6, 60.2)

46.6 (43.0, 50.2)

NPV

82.1 (81.6, 82.6)

81.7 (81.2, 82.3)

88.5 (85.5, 90.9)

87.6 (86.1, 89.0)

58.4 (53.7, 63.0)

78.6 (77.7, 79.4)

84.0 (82.9, 85.1)

89.9 (86.9, 92.3)

  1. Note: All estimates are expressed as %. PPV = positive predictive value, NPV = negative predictive value, AB Alberta, BC British Columbia, MB Manitoba, ON Ontario, QC Quebec, CPRD UK Clinical Practice Research Datalink. Highest and lowest site-specific estimates for each measure are delineated by italics and underline font, respectively